Zhao, H., Mayzus, R., Sun, S., Samimi, M., Schulz, J. K., Azar, Y., Wang, K., Wong, G. N., Gutierrez, Jr., F., Rappaport, T.S., 28 GHz Millimeter Wave Cellular Communication Measurements for Reflection and Penetration Loss in and around

Buildings in New York City, to appear in the 2013 IEEE International Conference on Communications (ICC), June 9∼13,2013.28 GHz Millimeter Wave Cellular Communication

Measurements for Reflection and Penetration Lossin and around Buildings in New York City

Hang Zhao, Rimma Mayzus, Shu Sun, Mathew Samimi, Jocelyn K. Schulz,Yaniv Azar, Kevin Wang, George N. Wong, Felix Gutierrez, Jr., Theodore S. Rappaport

NYU WIRELESSPolytechnic Institute of New York University, Brooklyn, NY 11201

tsr@nyu.edu

Abstract—In this paper, we present reflection coefficients andpenetration losses for common building materials at 28 GHzfor the design and deployment of future millimeter wave mobilecommunication networks. Reflections from walls and buildingsand penetration losses were measured for indoor and outdoormaterials, such as tinted glass, clear glass, brick, concrete, anddrywall at 28 GHz in New York City. A 400 Mega-chip-per-second sliding correlator channel sounder and 24.5 dBi steerablehorn antennas were used to emulate future mobile devices withadaptive antennas that will likely be used in future millimeterwave cellular systems [1]. Measurements in and around buildingsshow that outdoor building materials are excellent reflectors withthe largest measured reflection coefficient of 0.896 for tinted glassas compared to indoor building materials that are less reflective.We also found that penetration loss is dependent not only onthe number of obstructions and distance between transmitterand receiver, but also on the surrounding environment. Thegreatest penetration loss containing three interior walls of anoffice building was found to be 45.1 dB, with 11.39 m separationbetween the transmitter and receiver.

Index Terms—28 GHz, mm-wave communication, indoor-to-outdoor penetration, reflection, in-building propagation, buildingpenetration, 5G

I. INTRODUCTION

In recent years, the demand for high speed cellular dataand the need for more spectrum have motivated the use ofmillimeter wave (mm-wave) carrier frequencies for futurecellular networks, where high gain adaptive antennas, MIMO,and spatial beamforming can be implemented in very smallform factors [1][2][3][4]. Unused or underutilized LMDSbroadband spectrum exists at 28 GHz [4], and given thelow atmospheric absorption as compared to 60 GHz, thespectrum at 28 GHz has very comparable free space path lossas today’s 1-2 GHz cellular bands [2]. In addition, despitemyths to the contrary, rain attenuation and oxygen loss doesnot significantly increase at 28 GHz, and, in fact, may offerbetter propagation conditions as compared to today’s cellularnetworks when one considers the availability of high gainadaptive antennas and cell sizes on the order of 200 meters[4][5][6].

The design and optimization of future mm-wave mobilecommunication systems is highly influenced by the spatial

multipath characteristics of the wireless channel. A funda-mental relationship was established between the angular dis-tribution of power in a multipath channel and the narrowbandfading in a local region [7]. In addition, the performanceof Code Division Multiple Access (CDMA) cellular radiosystems in multipath environments was shown to improvedramatically with proper spatial filtering provided by adaptiveantenna arrays and switched beam systems [8]. These andother results from the 1990s offer promise for future beam-forming methodologies that may use narrowband pilot signalsto rapidly find, lock on, and combine the best spatial pathsusing highly directional adaptive antennas [6].

Besides the spatial parameters of the wireless channel,reflection and penetration characteristics of common buildingmaterials will also be required for the planning and design offuture mm-wave wireless communication systems.

Previous research showed that 9.6 GHz, 28.8 GHz and57.6 GHz waves propagate through a glass wall with virtuallyno loss, while the penetration loss increased by 25 dB to50 dB when the glass surface was metal coated [9]. Today,external building glass is routinely infused or coated withmetal to block ultraviolet rays while improving thermal insu-lating properties. In [9], no signal was detected through brickpillars throughout the mm-wave bands, yet signals through ahollow plasterboard wall resulted in a penetration loss rangingbetween 5.4 dB and 8.1 dB. In [10], researchers measuredpenetration losses of 2 dB, 9 dB, and 35.5 dB at 60 GHzthrough a glass door, a plasterboard wall with metallic studs,and a wall with a metal-backed blackboard, respectively [10].

The reflected multipath distribution was found to be highlycorrelated with the propagation environment at 60 GHz [11].Multipath reflections from buildings and other objects in anurban environment at street level, as measured with narrowbeam antennas, caused at least 15 dB additional path loss rel-ative to line of sight (LOS) signals [9]. In addition, 15 to 20 dBof attenuation in single-bounce reflected paths, as comparedto LOS paths, were measured in an office environment at 60GHz [12].

This work presents 28 GHz propagation measurementsin and around buildings in New York City. By comparing

transmitted power to reflected power measured from a materialor obstruction under test, the magnitude of the reflectioncoefficient |Γ||| for a vertically polarized plane wave can becomputed from (1) [13][14]:

|Γ||| =

√|Er|2120π Aε√|Ei|2120πAε

=

√|Er|2√|Ei|2

=|Er||Ei|

(1)

where |Er| and |Ei| are the magnitudes of the reflected andincident electric fields in V/m, respectively, and Ae is theantenna effective area measured in the far-field. In our case,a far field close-in reference distance was set at do = 5m for24.5 dBi antennas.

Penetration loss was calculated by comparing the differencein received power between unobstructed free space measure-ments versus the test material obstructing the transmitter (TX)and receiver (RX). Received power at d ≥ do with respect toa reference distance, do, can be calculated using (2) [15]:

PR(d) = PT +GT +GR − 20log10(4πdoλ

)−ECL[dB] (2)

where PR and PT are the received and transmitted powersin dBm, GT and GR are the transmitter and receiver antennagains of 24.5 dBi each with 10◦ half power beamwidth, λ is thewavelength of the carrier wave (10.71 mm at 28 GHz), do is 5m, the far-field close-in reference distance between the TX andRX, and ECL is excess channel loss in dB beyond do. In ourmeasurements, the different PT values used were -8.55 dBm,+11.63 dBm and +21.37 dBm. The corresponding EIRPs were+15.95 dBm, +36.13 dBm, and +45.87 dBm, respectively.

This paper is structured as follows: Section II describes theexperimental procedure and the equipment setup for reflectionand penetration measurements for common building materials.Section III presents the analysis for reflectivity and penetrationloss for different environments in New York City. Section IVconcludes with key results presented in this paper.

II. 28 GHZ PENETRATION AND REFLECTIONMEASUREMENT PROCEDURE

Throughout the measurement campaign conducted duringthe summer of 2012, reflection and penetration measurementswere recorded in and around many buildings on the NYUcampuses in Manhattan and Brooklyn. These sites includethe 10th floor of 2 MetroTech Center (MTC) at NYU-Polyin Brooklyn, Othmer Residence Hall (ORH) in Brooklyn,and Warren Weaver Hall (WWH) in Manhattan. Propagationcharacteristics in these urban environments were measuredusing the same broadband channel sounding hardware usedin [4][5][6][14]. When transmitting with full power (PT =+30 dBm before TX antenna), the largest path loss that ourhardware could have recorded was 178 dB (103 dB relativeto 5 meter free space). In this paper, all presented path lossesare in excess of the 5 meter free space reference path loss of75.3 dB at 28 GHz.

Fig. 1. Test setup at ORH and WWH for measuring reflected power fromvarious building materials at 28 GHz. The TX and RX were placed 2.5 maway from the material, with the antennas at equal heights at 1.5 m, orientedat two different incident angles, θ = 10◦ and 45◦. Both the TX and RX hornantennas had 24.5 dBi gains with 10◦ half power beamwidth.

Fig. 2. Test setup at ORH and WWH for measuring penetration loss ofvarious building materials at 28 GHz. To measure the penetration loss for atest material, first, the TX and RX were placed 5 m away from each other infree space (left). Then, they were moved and positioned at normal incidenceon opposite sides of test materials at equal heights of 1.5 m, with antennaspointing at each other, and both 2.5 m away from the test material (right).Both of the horn antennas had 24.5 dBi gains with 10◦ half power beamwidth.

To measure the reflected power from a test material, theTX and RX were separated by a distance of 2.5 m from thematerial with the antennas oriented at two different incidentangles, θ = 10◦ and 45◦, for tinted glass, clear glass, concrete,and drywall, as shown in Fig. 1.

As illustrated in Fig. 2, in order to conduct the penetrationloss measurement, the TX and RX were first separated by 5meters in free space conditions to provide a far-field referencepath loss, which was consistently measured to be 75.3 dB.Then, the TX and RX were moved to be on each side of atest material while still maintaining a 5 m separation distance.Both TX and RX were placed 2.5 m away from the testmaterial with identical antenna heights of 1.5 m as seen inFig. 2. In all tests, vertically polarized 28 GHz radio waveswere transmitted with directional antennas pointing at eachother, normal to the test material. Measurements of penetrationloss for common building materials (tinted glass, clear glass,brick, and drywall) were conducted at all three locations(ORH, MTC, and WWH). Fig. 3 shows the TX and RXsetup for outdoor reflection, and both indoor and outdoorpenetration measurements. The images are representatives of

Fig. 3. Pictures of the 28 GHz reflection measurement for outdoor tintedglass at ORH (top left), outdoor concrete wall at ORH (top right), penetrationloss measurement for indoor clear glass at MTC (bottom left) and tinted glassat ORH (bottom right).

the configuration used for all materials.At MTC, indoor measurements of penetration losses for

multiple walls and obstructions were conducted at eight dif-ferent RX locations as illustrated in Fig. 4. For RX 1 and2, the signal was transmitted through two and three walls,respectively. All of the walls measured here were made ofgypsum drywall and had the same thickness of 13 cm. ForRX 3 and 4, the signal was transmitted first through threewalls and then one door, and through three walls followedby two doors, respectively. The doors mentioned above weremade of wood with a thickness of 4.5 cm, and had a 60 cm ×10.2 cm glass window pane with a 2 mm-thick square-shapedmetal grating contained in a metal frame. For RX 5 and RX6, the signal propagated through two office cubicles and fourwalls successively. For RX 7, the signal went through twooffice cubicles, two walls, one wooden door, one wall, onemetal door, and two other walls in succession. For RX 8, thesignal was transmitted through two office cubicles, two walls,an elevator bank, and another three walls successively.

III. 28 GHZ REFLECTION AND PENETRATION ANALYSIS

Table I summarizes the reflection coefficients for commonbuilding materials at 28 GHz. The table compares results ofindoor materials to outdoor materials. Note that the outdoormaterials have larger reflection coefficients of 0.896 for tintedglass and 0.815 for concrete at a 10◦ incident angle, ascompared to clear non-tinted glass and drywall, which havelower reflection coefficients of 0.740 and 0.704, respectively.This result follows theory [15], and is most likely due tooutdoor building materials containing thick and dense metallayers. Reflective surfaces at 28 GHz allow radio frequency(RF) energy to be contained within the building, which reducesco-channel intercell interference that could leak outside of

Fig. 4. Indoor office wall penetration measurement plan at MTC 10th floor.The TX location is marked by a yellow star, the RX locations where signalscan be acquired are marked by green circles, and locations where weak signalscan be detected are red triangles. An outage (penetration loss is greater than74 dB) is shown as a black cross. The paths for wave penetration are shownin black arrows.

TABLE ICOMPARISON OF REFLECTIVITY FOR DIFFERENT MATERIALS AT 28 GHZ.CONCRETE AND DRYWALL MEASUREMENTS WERE CONDUCTED WITH 10◦

AND 45◦ INCIDENT ANGLES, WHILE TINTED AND CLEAR GLASSREFLECTIVITY WERE MEASURED AT 10◦ . BOTH OF THE HORN ANTENNAS

HAVE 24.5 DBI GAINS WITH 10◦ HALF POWER BEAMWIDTH.

the building, and thus suggests high frequency reuse betweenindoor and outdoor networks. This also illuminates difficultiesin achieving indoor coverage with outdoor infrastructure, asthe buildings appear very difficult to penetrate from the outsideat 28 GHz. Relays and ultrawideband repeaters will likelybe required to achieve indoor-to-outdoor coverage, or elseoutdoor mobile users will need to handoff into the indoornetwork (perhaps unlicensed spectrum or reused mm-wavespectrum) as a user enters a building.

Table II contains the penetration losses of building materialsat a 5 m TX-RX separation distance. The results for outdoortinted glass in ORH consistently show that a large portionof the signal power (|Γ| = 0.896) is reflected and couldnot penetrate through the glass. Brick also shows a highpenetration loss of 28.3 dB at 1.83 m thickness through a brickpillar. Indoor drywall has moderate attenuation of 6.84 dB andclear, non-tinted glass had the least attenuation of 3.6-3.9 dB.High penetration loss through brick and tinted glass, and low

TABLE IICOMPARISON OF PENETRATION LOSSES FOR DIFFERENT ENVIRONMENTS

AT 28 GHZ. THICKNESSES OF DIFFERENT COMMON BUILDING MATERIALSARE LISTED. BOTH OF THE HORN ANTENNAS HAVE 24.5 DBI GAINS WITH

10◦ HALF POWER BEAMWIDTH.

attenuation through clear glass and drywall suggest that RFenergy can be contained in intended areas within buildings,also reducing interference, but making building penetrationmore difficult.

To further examine the wave propagation during the pen-etration measurements at ORH, Fig. 5 shows a Power DelayProfile (PDP) measuring the penetration loss through the tintedglass. After initial penetration and reception (excess delay of0 ns) through the glass, the transmitted wave reflected withinthe building and was received with excess delays of 35 ns,50 ns, and 75 ns. We computed a penetration loss of 40.1dB through the tinted glass at ORH with Equation 2, wherereceived power was -75 dBm and 5 m free space receivedpower was -34.9 dBm, using a transmitted power of -8.55dBm and high gain 24.5 dBi horn antennas. Despite 40.1 dBpenetration loss through the tinted glass, the radiated signal

Fig. 5. Multipath delay spread observed in the 5 m tinted glass penetrationtest at ORH at 28 GHz. PR is the received power of the first arriving peak.It is also the multipath observed with the largest excess delay (75 ns) amongall the data collected for penetration measurements. The TX is positioned inan outdoor environment, and the RX is positioned indoors in a residentialenvironment. Both of the horn antennas have 24.5 dBi gains with 10◦ halfpower beamwidth.

Fig. 6. Potential reflectors that created multipath propagation observed afterinitial penetration through tinted glass at ORH at 28 GHz. The reflectionsmeasured in Fig. 5, travelled 75 feet (e.g. 75 ns) after the first incoming wave(LOS component) was received.

still had enough power to be received up to 75 ns, whichindicates there are many reflective materials in the buildingthat will propagate well at 28 GHz, as shown in Fig. 6.

Table III presents the values for penetration losses throughmultiple interior rooms and walls in an office environment,with variations in TX-RX seperation distance and the numberof penetration obstructions as seen in Fig. 4. Data is groupedinto three subsections: signal acquired, signal detected, and nosignal detected, as shown in Fig. 4. Signal acquired is definedas a location where the SNR is sufficiently high for accurateacquisition, i.e. penetration loss (relative to a 5 meter freespace test) is less than 64 dB beyond the 5 m free space loss.Signal detected is a location where the SNR is high enoughto distinguish signal from noise; however, SNR is not largeenough to be accurately acquired, i.e. penetration loss between64 dB and 74 dB. No signal detected denotes an outage, wherepenetration loss is greater than 74 dB beyond the 5 m freespace loss.

As shown in Fig. 4 and Table III, the results obtained ateach RX site show that the penetration loss does not greatlydepend on the TX-RX separation distance. Although RX 5has a separation distance of 25.6 m (two cubicles and fourwalls of obstructions), which is 14.2 m greater than that ofRX 3 (three walls and a door of obstructions), both locationshave virtually identical measured penetration loss. Note thatamong the measured indoor RX sites, one outage was foundat RX 8 with the TX-RX separation distance of 35.8 m. Onepossible cause for this outage is that the separation distancewas too large and unable to penetrate the metal elevators inthe elevator bank.

TABLE IIIRESULTS FOR INDOOR WALL PENETRATION LOSS MEASUREMENTS IN ANOFFICE ENVIRONMENT AT 28 GHZ. THE RX ID NUMBERS CORRESPOND

TO THE LOCATIONS LABELED IN FIG. 4.RX LOCATIONS ARE ORDERED ININCREASING TX-RX SEPARATION DISTANCES. MULTIPLE OBSTRUCTIONSEXISTED BETWEEN THE TX AND RX. “SIGNAL DETECTED” ARE DENOTED

BY LOCATIONS WHERE THE SNR WAS HIGH ENOUGH TO DISTINGUISHSIGNAL FROM NOISE; HOWEVER, SNR WAS NOT LARGE ENOUGH FOR THE

PDP TO BE ACQUIRED. “NO SIGNAL DETECTED” DENOTES AN OUTAGEWHERE MATERIAL PENETRATION LOSS IS GREATER THAN 74 DB.

IV. CONCLUSION

This paper presents measurements of reflection and pen-etration of common building materials at 28 GHz, usingvertically polarized antennas in New York City. Utilizing abroadband sliding correlator channel sounder, received powerwas measured for various building materials, such as brick,concrete, drywall, tinted glass, and clear glass. Reflectioncoefficients for outdoor materials such as tinted glass andconcrete were as large as 0.896, showing high reflectivityand very small transmittivity. Outdoor tinted glass had apenetration loss of 40.1 dB, indicating that such material ishighly attenuative and very reflective. In contrast, indoor clearnon-tinted glass had a penetration loss of 3.9 dB. In addition,indoor walls had a penetration loss of 6.84 dB at 28 GHz. In adensely urban environment, brick penetration through a pillarcaused a 28.3 dB penetration loss. From the measurementspresented here, it is clear that indoor-to-outdoor penetrationwill be quite difficult at 28 GHz, whereas indoor-to-indoorand outdoor-to-outdoor propagation is easily supported bythe strong reflectivity of external building materials and lowattenuation of indoor materials. The highly reflective externalbuilding materials can enhance outdoor signal coverage, andallow a wide range of possible angles of arrival to createTX-RX links in outdoor environments. Further, it was foundthat the penetration loss through indoor walls is dependenton the number of obstructions and the characteristics of theobstructions and the TX-RX separation distance. In addition,the lower penetration loss of indoor materials coupled withreflective outdoor materials and highly lossy external glass andwalls helps reduce interference between indoor and outdoormm-wave networks, thus allowing high frequency reuse. Thesemeasurement results of common building materials providepractical values for designing future mm-wave broadbandcellular communication systems operating at 28 GHz.

V. ACKNOWLEDGMENT

This work was sponsored by Samsung DMC R&D Commu-nications Research Team (CRT) through Samsung Telecom-munications America, LLC. The authors wish to thank GeorgeMacCartney, Shuai Nie, and Junhong Zhang of NYU WIRE-LESS, as well as DuckDong Hwang and Shadi Abu-Surraof Samsung, NYU administration, NYU Public Safety, andNYPD for their contribution to this project. Measurementsrecorded under U.S. FCC Experimental License 0040-EX-ML-2012.

REFERENCES

[1] F. Gutierrez, S. Agarwal, K. Parrish, and T. S. Rappaport, “On-ChipIntegrated Antenna Structures in CMOS for 60 GHz WPAN systems,”IEEE Journal on Selected Areas in Communications, vol. 27, no. 8, pp.1367–1378, Oct. 2009.

[2] T. S. Rappaport, J. N. Murdock, and F. Gutierrez, “State of the Art in60-GHz Integrated Circuits and Systems for Wireless Communications,”Proceedings of the IEEE, vol. 99, no. 8, pp. 1390–1436, Aug. 2011.

[3] P. Zhouyue and F. Khan, “An Introduction to Millimeter-wave MobileBroadband Systems,” IEEE Communications Magazine, vol. 49, no. 6,pp. 101–107, June 2011.

[4] T. S. Rappaport, E. Ben-Dor, J. N. Murdock, and Y. Qiao, “38 GHzand 60 GHz Angle-dependent Propagation for Cellular and Peer-to-PeerWireless Communications,” in 2012 IEEE International Conference onCommunications (2012 ICC), June 10-15, 2012, Ottawa, Canada, pp.4568 –4573.

[5] E. Ben-Dor, T. S. Rappaport, and Y. Qiao, “Millimeter-wave 60 GHzOutdoor and Vehicle AOA Propagation Measurements using a Broad-band Channel Sounder,” in 2011 IEEE Global TelecommunicationsConference, Dec. 2011.

[6] Y. Azar and et al., “28 GHz Propagation Measurements for OutdoorCellular Communications Using Steerable Beam Antennas in New YorkCity,” IEEE International Conference on Communications (2013 ICC),to appear, June 9-13, 2013, Budapest, Hungary.

[7] G. Durgin and T. S. Rappaport, “Basic Relationship Between MultipathAngular Spread and Narrowband Fading in Wireless Channels,” IETElectronics Letters, vol. 34, no. 25, pp. 2431–2432, Dec. 10, 1998.

[8] J. C. Liberti and T. S. Rappaport, “Analysis of CDMA cellular radiosystems employing adaptive antennas in multipath environments,” inIEEE 46th Vehicular Technology Conference, vol. 2, April-May 1996,pp. 1076–1080.

[9] E. J. Violette, R. H. Espeland, R. O. DeBolt, and F. K. Schwering,“Millimeter-wave Propagation at Street Level in an Urban Environment,”IEEE Transactions on Geoscience and Remote Sensing, vol. 26, no. 3,pp. 368–380, May 1988.

[10] K. Sato, T. Manabe, T. Ihara, H. Saito, S. Ito, T. Tanaka, K. Sugai,N. Ohmi, Y. Murakami, M. Shibayama, Y. Konishi, and T. Kimura,“Measurements of Reflection and Transmission Characteristics of Inte-rior Structures of Office Building in the 60-GHz Band,” IEEE Trans-actions on Antennas and Propagation, vol. 45, no. 12, pp. 1783–1792,Dec. 1997.

[11] H. Xu, V. Kukshya, and T. S. Rappaport, “Spatial and TemporalCharacteristics of 60-GHz Indoor Channel,” IEEE Journal on SelectedAreas in Communications, vol. 20, no. 3, pp. 620–630, Apr. 2002.

[12] A. Malsev, R. Maslennikov, A. Sevastyanov, A. Khoryaev, and A. Lo-mayev, “Experimental Investigations of 60 GHz WLAN Systems in Of-fice Environment,” IEEE Journal on Selected Areas in Communications,vol. 27, no. 8, pp. 1488–1499, Oct. 2009.

[13] T. S. Rappaport, Wireless Communications: Principles and Practice,2nd ed. New Jersey: Prentice Hall, 2002.

[14] O. Landron, M. J. Feuerstein, and T. S. Rappaport, “In Situ MicrowaveReflection Coefficeint Measurements for Smooth and Rough ExteriorWall Surfaces,” in IEEE Vehicular Technology Conference (1993 VTC),May 18-20, 1993, Secaucus, NJ.

[15] G. Durgin, T. S. Rappaport, and H. Xu, “5.85-GHz Radio Path Lossand Penetration Loss Measurements in and Around Homes and Trees,”IEEE Communications Letters, vol. 2, no. 3, pp. 70–72, Mar. 1998.